Anthropic needs to convince the administration they can fix or implement new guardrails to reassure that the model is safe for access beyond US nationals. The pace of development in AI foundation models and their sophistication, particularly Mythos and Fable 5, has spooked both commercial and government worlds. Some threats from AI capabilities are raising questions about reliance on US-sourced models, potentially accelerating sovereign AI development globally.

OpenAI's spending of $34B on $20B revenue is sustainable in the short term because so much money is pouring into the AI market. Investors understand that to grow into the $200B+ revenue run rate by 2030, they need to continue investing in infrastructure. I wouldn't be surprised if projections anticipate deeper cash outflows in coming years as they build out that infrastructure.

NVIDIA's first debt raise taps into the broad capital markets enthusiasm for the AI trade, helped by the SpaceX IPO which was $25B and oversubscribed over 3 times. Every major US tech company has come to market in similar size this year. It gives NVIDIA a stronger capital structure, optionality for organic investment and acquisitions, and probably reduces their cost of capital.